SOUNIO V2 • EPISTEMIC COMPUTING

Compute at the boundary between certainty and consequence.

Sounio is a systems programming language for scientific teams that need to reason not only about values, but about trust, provenance, and uncertainty at every stage of execution.

Start with Docs Open Playground

Featured medical case study: Vancomycin Dosing That Cannot Lie About Uncertainty.

Built for teams shipping high-consequence scientific software.

Why Sounio

Modern scientific systems need mechanical sympathy and epistemic honesty in the same language runtime.

Epistemic Types

Every value can carry uncertainty, confidence, and provenance directly in the type system.

Systems Performance

Native backends, GPU kernels, and explicit effects deliver low-level control without trust blind spots.

Scientific Rigor

Built for domains where correctness is not enough and confidence quality drives real-world decisions.

Auditability

Traceability and evidence-oriented execution paths make scientific outputs defensible and reproducible.

The Sounio Difference

See how Sounio eliminates entire classes of scientific software bugs that traditional languages silently ignore.

Every measurement carries doubt. How does the language handle it?

Python + NumPy14 lines
import numpy as np

value = 9.8
std_dev = 0.3

# Manual propagation
derived = value * 2.5
derived_err = std_dev * 2.5  # hope this is right

# Manual threshold check
if derived_err / derived < 0.05:
    approve(derived)
else:
    # What was the original source?
    # Who measured it? When?
    remeasure()  # no provenance
  • Uncertainty is a separate variable — easy to lose
  • Propagation rules are manual and error-prone
  • No provenance: source identity is gone
Sounio11 lines
let c: Knowledge<mg_per_l> = measure(
  value: 9.8,
  uncertainty: 0.3,
  confidence: 0.97,
  source: "lab_run_042"
)

let derived = c * 2.5  // uncertainty auto-propagates

if derived.confidence > 0.95 {
  approve(derived)  // provenance travels with it
}
  • Uncertainty is part of the type — impossible to forget
  • Propagation is automatic through all operations
  • Provenance tracks the full measurement chain

Scientific Domains

Domain pipelines where confidence and traceability can change outcomes.

Pharmacometrics

PK/PD pipelines with confidence-aware dosing decisions.

Explore domain

Neuroimaging

fMRI and signal pipelines with provenance-aware thresholds.

Explore domain

Quantum

VQE and uncertainty propagation from noisy measurement loops.

Explore domain

Climate Science

Ensemble projections with explicit confidence gates.

Explore domain

Built For High-Consequence Scientific Systems

Sounio combines low-level systems control with explicit uncertainty and provenance semantics, so teams can ship faster without sacrificing auditability.

Sounio visual identity artwork

Verification Layer

Official identity assets anchor the trust model across docs, platform surfaces, and release channels, so users can verify they are reading canonical Sounio materials.

Sounio Language emblem badge

Current Platform Depth

  • 627K+Language + runtime lines
  • 236K+Standard library footprint
  • 15M+Ontology entities integrated
  • 6Locale routes supported

Ship scientific software that explains itself.

Build systems where every number can be questioned, traced, and trusted before it reaches a decision.

Enter the Learning Hub Explore on GitHub 
curl -sSf https://souniolang.org/install.sh | sh